Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when an analysis is complete and verified, and you need to decide how to report it and archive the work for reproducibility
.claude/skills/k-dense-ai-reporting-and-archiving-findings/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 146% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 111% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 33% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 20% | 0% |
| case-22 | ✗→✓ | ▲ Improved | 50% | 0% |
Complete an investigation by confirming reproducibility, presenting clear options, handling the chosen one, and archiving everything needed to reproduce the result.
Core principle: Confirm reproducibility -> separate confirmatory from exploratory -> present options -> execute choice -> archive code + data + environment + pre-registration.
Announce at start: "I'm using the reporting-and-archiving-findings skill to complete this work."
Before reporting anything, confirm the whole analysis reproduces from immutable raw data with the fixed seed in the pinned environment.
bash# From a clean state: re-run the pipeline end to end # Confirm headline numbers match what you intend to report
Use science-superpowers:verifying-results-before-claiming. If it doesn't reproduce, stop — fix reproducibility (possibly via science-superpowers:investigating-anomalous-results) before reporting. Don't report a number you can't regenerate.
For any confirmatory claim, also run the pre-registration audit and keep its output for the report:
bash# ships with science-superpowers:preregistering-analysis <skills root>/preregistering-analysis/prereg.sh audit
If the audit fails, the confirmatory label is indefensible — the registration changed after its freeze, or outputs predate it. Relabel the affected analyses exploratory (or pre-register a fresh test on unused data) before reporting.
bashGIT_DIR=$(cd "$(git rev-parse --git-dir)" 2>/dev/null && pwd -P) GIT_COMMON=$(cd "$(git rev-parse --git-common-dir)" 2>/dev/null && pwd -P) BRANCH=$(git branch --show-current)
GIT_DIR == GIT_COMMON: normal repo, no worktree cleanup neededGIT_DIR != GIT_COMMON, named branch: worktree, provenance-based cleanup (Step 5)bashgit merge-base HEAD main 2>/dev/null || git merge-base HEAD master 2>/dev/null
Or ask: "This branch split from main — correct?"
Normal repo / named-branch worktree — present exactly these 4 options:
Analysis complete and reproducible. What would you like to do?
1. Merge the analysis back to <base-branch> locally
2. Write up and share (report / preprint / pull request)
3. Keep the branch as-is (I'll handle it later)
4. Discard this work
Which option?Detached HEAD — present these 3 (no local merge):
Analysis complete and reproducible (externally managed workspace).
1. Push as a new branch and open a pull request / share
2. Keep as-is
3. Discard this work
Which option?Don't add explanation — keep options concise.
bashMAIN_ROOT=$(git -C "$(git rev-parse --git-common-dir)/.." rev-parse --show-toplevel) cd "$MAIN_ROOT" git checkout <base-branch> && git pull && git merge <feature-branch> # Re-run the pipeline on the merged result; confirm it still reproduces
Then cleanup worktree (Step 6), then git branch -d <feature-branch>.
Produce the report (see "Report Content" below). If sharing via PR:
bashgit push -u origin <feature-branch> gh pr create --title "<title>" --body "$(cat <<'EOF' ## Question <the research question> ## What was done <the pre-registered analysis, and any documented deviations> ## Findings <confirmatory results with effect sizes + intervals> ## Exploratory (not confirmatory) <clearly separated leads> ## Reproducibility - Pre-registration: <path/commit> - Environment: <lockfile> - Seed: <value> - Re-run: <command> EOF )"
Do NOT clean up the worktree — it's needed for iteration on feedback.
Report: "Keeping branch <name>. Worktree preserved at <path>." No cleanup.
Confirm first:
This will permanently delete:
- Branch <name>
- All commits: <list>
- Worktree at <path>
Type 'discard' to confirm.Wait for the exact word. Then cleanup worktree (Step 6) and git branch -D <feature-branch>.
Only for Options 1 and 4. Options 2 and 3 preserve the worktree.
bashWORKTREE_PATH=$(git rev-parse --show-toplevel)
GIT_DIR == GIT_COMMON: normal repo, nothing to clean up..worktrees/, worktrees/, or ~/.config/superpowers/worktrees/: we own it.bash MAIN_ROOT=$(git -C "$(git rev-parse --git-common-dir)/.." rev-parse --show-toplevel) cd "$MAIN_ROOT" git worktree remove "$WORKTREE_PATH" git worktree prune
Every report MUST:
Whatever the option, ensure the archive contains everything needed to regenerate the result:
prereg.sh audit outputNever:
Always:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 13,280 | 29,874 | +125% | 1 | 1 | 0% | 2,208 | 2,024 | -8% | 0 | 0 | — |
case-06 | pass→pass | 9,602 | 5,699 | -41% | 1 | 1 | 0% | 1,744 | 2,697 | +55% | 0 | 0 | — |
case-02 | fail→fail | 3,681 | 2,778 | -25% | 1 | 1 | 0% | 171 | 2,053 | +1101% | 0 | 0 | — |
case-03 | fail→fail | 2,712 | 2,563 | -5% | 1 | 1 | 0% | 370 | 2,031 | +449% | 0 | 0 | — |
case-04 | pass→pass | 15,222 | 5,226 | -66% | 1 | 1 | 0% | 2,325 | 2,579 | +11% | 0 | 0 | — |
case-05 | pass→pass | 12,162 | 5,600 | -54% | 1 | 1 | 0% | 1,951 | 2,695 | +38% | 0 | 0 | — |
case-07 | pass→pass | 8,514 | 2,497 | -71% | 1 | 1 | 0% | 1,358 | 2,071 | +53% | 0 | 0 | — |
case-08 | fail→pass | 5,416 | 2,045 | -62% | 1 | 1 | 0% | 787 | 1,933 | +146% | 0 | 0 | — |
case-09 | pass→pass | 4,233 | 3,514 | -17% | 1 | 1 | 0% | 666 | 2,257 | +239% | 0 | 0 | — |
case-10 | fail→pass | 9,157 | 9,461 | +3% | 1 | 1 | 0% | 1,467 | 3,098 | +111% | 0 | 0 | — |
case-11 | pass→pass | 10,277 | 8,340 | -19% | 1 | 1 | 0% | 1,673 | 3,121 | +87% | 0 | 0 | — |
case-12 | fail→pass | 14,571 | 9,062 | -38% | 1 | 1 | 0% | 2,219 | 2,960 | +33% | 0 | 0 | — |
case-13 | pass→pass | 10,425 | 2,861 | -73% | 1 | 1 | 0% | 1,768 | 2,141 | +21% | 0 | 0 | — |
case-14 | pass→pass | 11,913 | 16,370 | +37% | 1 | 1 | 0% | 2,104 | 4,373 | +108% | 0 | 0 | — |
case-15 | pass→pass | 12,708 | 19,888 | +56% | 1 | 1 | 0% | 2,200 | 5,153 | +134% | 0 | 0 | — |
case-16 | pass→pass | 20,128 | 18,355 | -9% | 1 | 1 | 0% | 3,161 | 4,612 | +46% | 0 | 0 | — |
case-17 | pass→pass | 8,644 | 4,841 | -44% | 1 | 1 | 0% | 1,320 | 2,451 | +86% | 0 | 0 | — |
case-18 | pass→pass | 6,378 | 2,884 | -55% | 1 | 1 | 0% | 984 | 2,099 | +113% | 0 | 0 | — |
case-19 | fail→pass | 13,126 | 4,700 | -64% | 1 | 1 | 0% | 2,084 | 2,499 | +20% | 0 | 0 | — |
case-20 | pass→pass | 8,552 | 2,594 | -70% | 1 | 1 | 0% | 1,471 | 2,096 | +42% | 0 | 0 | — |
case-21 | pass→pass | 12,488 | 6,948 | -44% | 1 | 1 | 0% | 1,845 | 2,736 | +48% | 0 | 0 | — |
case-22 | fail→pass | 13,234 | 8,628 | -35% | 1 | 1 | 0% | 1,966 | 2,944 | +50% | 0 | 0 | — |
case-23 | pass→pass | 10,621 | 7,585 | -29% | 1 | 1 | 0% | 1,567 | 2,819 | +80% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted, and 22 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +22 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.